8th Baltic Congress of Radiology

Radioprotection and management

poster abstract

Nowadays there are multiple methods for patient dose registration including automatic and manual methods. Automatic dose analysis software collects the dose data from dose report in PACS server or X-ray machine and stores the dose data. Manual patient dose registration is managed by radiographer - patient dose is read from the screen of x-ray machine and written in the examination journal or hospital or radiology information system (RIS). Most of the hospitals in Latvia uses manual dose registration method that may cause different registration errors that could affect the dose analysis. This study was performed to determine the impact of dose registration errors on dose analysis.

Two dose analysis were carried out for the same CT patient examinations performed in Riga East university hospital in 2021. More than 8 000 patient exams were included in this study. For the first dose analysis manually collected data from radiology information system (RIS) was used and for the second dose analysis data from automatic dose monitoring software DoseWatch by General Electric was used. 4 CT machines and 2 different exam types (head CT and thorax CT) were included in this study. The median DLP dose (mGy*cm2) was calculated for each exam type and each CT unit from two different data sets. The obtained median dose values from two different data sets were compared and the difference was evaluated. In addition calculated median dose values were compared with National DRL standardlevel.

In thorax examination median dose calculated from RIS data set was underestimated in comparison with data obtained from DoseWatch for all 4 CT (-10%, -20%, -20% and -12%). The mean difference in thorax examinations was -15%. In head examination median dose calculated from RIS data was overestimated for 2 CT machines (1% and 33%) and underestimated for 2 CT (-6% and -15%). The highest difference for median dose in two different data sets was for CT2 in head examinations (33%). In 6 cases the median dose calculated from RIS data set was underestimated and in 2 cases it was overestimated. In comparison with dose standardlevels for head CT median dose exceeded the standardlevel in one CT machine for both data sets. For thorax examination median dose exceeded the sandardlevel for three CT machines for data set obtained in automatic dose management software and in 2 CT machines for data set obtained from the manual dose entry in the RIS.

From the study it was concluded that median dose varies significantly depending on a data set therefore manual entry errors significantly affect the result of the dose analysis. In most cases median dose obtained from manual dose registration method (RIS) was underestimated. It was concluded that automatic dose management softwares should be preferred for dose analysis. In case automatic dose management software is not available in the hospital, the manual dose registration should be evaluated with caution and should be revised precisely.

The aim was to determine the impact of dose registration errors on dose analysis. Median dose was calculated for the same CT patient examinations from two data sets (manually and automatically collected data). In 6 cases the median dose calculated from manually collected data set was underestimated and in 2 cases it was overestimated. It was concluded that manual entry errors significantly affect the result and automatic dose management softwares should be preferred.

1225